Hacker News Reader

Just watched the video, but is it correct to assume the inputs to this are:

- Manually specifying layers in the image by depth.

- Manually specifying all the objects you'd want to animate.

- Providing an animation direction and type.

Then it generates an animation?

From the abstract[1], the only input is selecting a few of the repeating elements, and drawing motion lines. Depth and background are detected automatically.

[1] https://pixl.cs.princeton.edu/pubs/Willett_2018_AMI/

Reply on news.ycombinator.com